Table 2 Proportion of phenotypic variance captured by all DNAm sites \((\hat \rho ^2)\) and phenotypic variance \((\hat \sigma _P^2)\) estimated from different OREML models.

From: Significant out-of-sample classification from methylation profile scoring for amyotrophic lateral sclerosis

OREML model (y = Cβ + Wu + e)

\(\hat \rho ^2\) (s.e.)

\(\hat \sigma _P^2\) (s.e.)

No covariates

15% (0.05)

0.247 (0.01)

With predicted age + predicted smoking score + sex + batch effects

17% (0.06)

0.227 (0.01)

With predicted cell proportions

24% (0.06)

0.237 (0.01)

With predicted age + predicted smoking score + sex + batch effects + predicted cell proportions

31% (0.08)

0.221 (0.01)

  1. s.e. standard error. In the absence of covariates \(\hat \sigma _P^2\) = P(1 − P), where P is the proportion of the sample that are cases.
Back to article page